Self Sufficient Precision Agriculture Management for Irrigation
With the growing adoption of the Internet of Things (IoT), connected devices have penetrated every aspect of our life, from health and fitness, home automation, automotive and logistics, to smart cities and industrial IoT.
Thus, it is only reasonable that IoT, connected devices, and automation would find its application in agriculture, and as such, tremendously improve nearly every facet of it.
Farming has seen a number of technological transformations in the last decades, becoming more industrialized and technology-driven. By using various smart agriculture gadgets, farmers have gained better control over the process of raising livestock and growing crops, making it more predictable and improving its efficiency.
Additionally, With 10% of all enterprise-generated data being generated and analyzed outside of the cloud or centralized data centers, edge computing is slowly gaining momentum across many industries.
However, as the figure is expected to reach 75% by 2025, we can soon witness exponential growth in edge computing adoption.
Namely, edge computing has already become a major IoT trend in agriculture, and for a good reason: edge computing wins hands down in terms of speed and efficiency when compared to the cloud infrastructure.
In this article, we will explore one of the IoT edge computing use case in agriculture as it is implemented in LightKone project. We apply this LightKone use case for irrigation management in Subsurface Drop Irrigation method.
Till now in the traditional way the water is pumped out from the well (or other water source) and transmitted to the polymer tubes. The water usually is used to irrigate multiple farms. However, every farm has different characteristics (soil, area, etc.) and the irrigation should be adapted taking account these parameters. For example, other piece of land needs more water and other piece of land needs less water. In order to avoid under-watering, the farmers usually irrigate more time than it is necessary. Problem: Water waste, energy waste (electricity for the pump), drainage problem (since the same time many farmers irrigating and the water in the underground water dump is not enough for everyone).
The irrigation procedure is taking placing following only empirical and observation rules. E.g. we irrigate for 2 hours and then we repeat every 10 days in the summer period. Sometimes either the citrus trees need more water because of extended high temperatures or less water due to lower temperatures.
In general, one of the main drawbacks of the SDI systemsis that water applications may be largely unseen, and it is more difficult to evaluate system operation and water application uniformity. System mismanagement can lead to under irrigation, less crop yield quality reductions, and over irrigation. The last may result among the others in poor soil aeration and deep percolation problems.
In the current situationwe don’t have a clue which part of the farm is either over-watered or under-watered. We can see only afterwards when the tree is not full of products or its leaves are yellow, etc. (again empirical and observation methods). A non-proper irrigation could affect 25–30% of the annual production.
Using the Self-Sufficient management system, the farm could be divided in clusters (e.g. as installed the polymer tubes) and accordingly will be distributed the LightKone self-sufficient nodes, containing the management unit, sensors and actuators. In that way the farmer could divide into zones his farms and when a zone is sufficient irrigated (retrieved value from the sensors) the actuators will stop the water flow into specific parts of the tubes. The rest part of the farm that still needs water will be still irrigated.
We will use LightKone technology to provide reliable computation and communication ability despite unreliable nodes and communication. We will present a Proof of concept using Lasp-on-GRiSP and Yggdrasil. Lasp provides a reliable replicated key/value store that runs with very little computational resources, on top of a communication layer, Partisan, that ensures reliable communication despite highly unreliable connectivity (using hybrid gossip). Basic connectivity provided by Yggdrasil underneath Partisan. We will extend Lasp with a simple task model that stores the management software in the Lasp store itself (which is possible because of higher-order nature of Erlang), and performs periodic computations, storing results in the Lasp store. GRiSP provides native Erlang functionality running with low power, with basic processor power and memory and wireless connectivity. GRiSP also provides Pmod sensor interfacing to provide the sensor and actuator abilities.
GRiSP nodes can be powered by solar batteries. 100% uptime is not required because of the Lasp redundancy. Occasional problems in individual nodes are solvable by periodic reboot of individual nodes. This will not hinder overall system operation because Lasp replication and Partisan hybrid gossip are designed to survive such problems.
Management policy control is provided by a connection to the sensor array, either by PC or cloud, which the farmer can do at any time. This connection does not need to be continuous or reliable. The management will continue to work even if the connection is not done for several days or more.
In that sense the self-sufficient system couldin a low-costmanner allow the farmer seamlessly to perform irrigation as needed by receiving values from sensors in the soil and controlling actuators to start/stop the irrigation procedure to specific rows of the tubes and accordingly part/cluster of the farm.
By Giorgos Kostopoulos, Gluk Advice B.V.